All checks were successful
deploy / deploy (push) Successful in 31s
Proof: Open the next implementation turn focused on durable downstream outcome linkage and settled inventory attribution for executed quotes. Assumptions: I017 and I018 belong to one coherent proof topic for the active NEAR Intents BTC/EURe path. Still fake: Planning defines the next truth gap and validation target, but no downstream outcome or settlement attribution implementation exists yet.
138 lines
7.2 KiB
Markdown
138 lines
7.2 KiB
Markdown
# Implementation Proof: execution outcome truth and settled attribution
|
|
|
|
Status: open
|
|
Opened: 2026-04-09
|
|
|
|
## Target outcome
|
|
This turn proves that `unrip` can move recent quote rows past submission-only evidence and into durable downstream outcome truth.
|
|
|
|
For the live NEAR Intents BTC/EURe system, the operator must be able to answer:
|
|
- was the quote merely submitted
|
|
- did the venue later complete or not fill it
|
|
- what settled inventory delta, if any, was actually observed afterward
|
|
- what part of the post-trade change is still inferred versus durably linked
|
|
|
|
## Why this is a meaningful architecture test
|
|
The previous turn fixed an overclaiming dashboard, but it also exposed the next hard truth gap:
|
|
- `submitted` is now shown truthfully as non-terminal evidence
|
|
- downstream completion and non-fill states are still mostly absent from durable repo-owned data
|
|
- recent asset-term tables still describe submitted quote terms, not settled inventory truth
|
|
- realized attribution is still fake because quote response success is not the same thing as completed asset movement
|
|
|
|
That means the system is now honest about the gap, but the gap still exists. This turn is where the shared history path must become more real again.
|
|
|
|
## Hypothesis
|
|
`unrip` becomes materially more trustworthy if it durably links:
|
|
- quote
|
|
- strategy decision
|
|
- execute command
|
|
- executor submission result
|
|
- downstream venue or relay outcome where available
|
|
- later settled inventory deltas attributable to that quote
|
|
|
|
The turn passes only if the operator can distinguish submission from later outcome and can inspect a recent completed path with explicit settled attribution instead of implied completion.
|
|
|
|
## Scope
|
|
- [I017] Per-trade realized attribution: link each successful trade to actual settled inventory deltas and attributable costs so the system can report realized net contribution per trade instead of only expected edge.
|
|
- [I018] Quote-to-outcome and markout analytics: link quote, response, decision, execution result, and settlement, then store later reference-price markouts to measure adverse selection and quote quality.
|
|
- Focus on the active NEAR Intents BTC/EURe venue and the repo-owned path already in production.
|
|
- Cover recent rows first; historical backfill is secondary unless required to prove the live path.
|
|
|
|
## Assumptions
|
|
- The venue or relay emits enough downstream identifiers or status surfaces to link at least some submitted quotes to later terminal outcomes.
|
|
- Inventory snapshots and durable history can support quote-to-settlement attribution for the active pair, even if the first version uses a narrow matching heuristic.
|
|
- Some rows will still remain unattributed or partially inferred; the UI and stored records must say that plainly.
|
|
- This turn should prioritize truthful linkage and settlement evidence over broad analytics polish.
|
|
|
|
## Turn-shaping rules
|
|
- `submitted` remains non-terminal evidence.
|
|
- `completed`, `not filled`, `settled`, `realized`, and `asset delta` are forbidden unless backed by a durable linked record representing that fact.
|
|
- If settlement attribution is heuristic rather than exact, the stored attribution record and operator surface must say so explicitly.
|
|
- Do not build a broad analytics suite before the quote-to-outcome path is durably linked.
|
|
- Prefer one repo-owned outcome and attribution model reused by storage, dashboard, and later analytics.
|
|
|
|
## Non-goals
|
|
- No new venue integration.
|
|
- No generalized PnL accounting for all treasury cashflows.
|
|
- No broad historical backfill project unless needed to prove current-path linkage.
|
|
- No strategy changes beyond what is required to preserve truthful lifecycle and attribution semantics.
|
|
|
|
## Required operator behavior
|
|
|
|
### Outcome truth
|
|
For a recent submitted quote, the dashboard must be able to render one of:
|
|
- `Submitted`
|
|
- `Awaiting outcome`
|
|
- `Not filled`
|
|
- `Completed`
|
|
|
|
The label must be driven by the strongest durable evidence currently stored.
|
|
|
|
### Settlement truth
|
|
For a recent completed quote, the operator must be able to inspect:
|
|
- quote id
|
|
- decision id
|
|
- command id
|
|
- executor result id or submission timestamp
|
|
- linked terminal outcome evidence
|
|
- linked settled asset delta or explicitly missing settled delta
|
|
|
|
### Attribution truth
|
|
If the system reports realized contribution, it must be derived from:
|
|
- linked settled asset movement
|
|
- explicit attributable costs where available
|
|
|
|
If costs or full settlement are missing, the system must not claim net realized truth.
|
|
|
|
### Markout truth
|
|
If markout is stored for completed or submitted rows, it must:
|
|
- be clearly labeled as later reference-price comparison
|
|
- remain separate from realized settlement attribution
|
|
|
|
## Semantic invariants
|
|
The implementation and tests must enforce at least:
|
|
- `submitted != completed`
|
|
- `completed` requires linked terminal outcome evidence
|
|
- settled asset delta requires linked settlement evidence, not quote terms
|
|
- markout is not realized PnL
|
|
- unattributed inventory movement must not be silently assigned to a quote
|
|
|
|
## Definition of done
|
|
- A durable quote outcome link exists from submission to later outcome where available.
|
|
- The repo stores an explicit per-quote outcome or attribution record for the active path.
|
|
- The dashboard can show at least one recent path as `Completed` only when durable linked evidence supports it.
|
|
- The dashboard can show non-fill or still-awaiting paths truthfully where that evidence exists.
|
|
- Recent settled asset delta or explicit lack of attribution is visible per completed row.
|
|
- Submission-term tables and summaries do not overclaim settled truth.
|
|
- Regression tests cover the negative semantic boundaries above.
|
|
|
|
## Validation evidence required
|
|
- direct evidence from deployed dashboard or bootstrap payload that recent rows separate `Submitted`, `Blocked`, `Rejected`, `Not filled`, and `Completed` when the evidence exists
|
|
- direct evidence that a completed row shows linked settled attribution rather than quote terms alone
|
|
- direct evidence that a submission-only row still does not claim completion or realized asset movement
|
|
- automated test evidence for quote-to-outcome linkage
|
|
- automated test evidence for settlement attribution boundaries
|
|
|
|
## Failure conditions
|
|
- the system still cannot represent downstream terminal outcome beyond submission
|
|
- any UI or backend surface claims completion from submission-only evidence
|
|
- settled asset delta is still derived from quoted terms rather than linked settlement evidence
|
|
- markout or later reference comparison is presented as realized PnL
|
|
- outcome attribution silently guesses across ambiguous rows without recording uncertainty
|
|
|
|
## Current real before this turn
|
|
- quote, decision, command, and execution submission result are durable
|
|
- the dashboard now renders those stages truthfully
|
|
- submission counts and recent submission terms are constrained to submission evidence
|
|
|
|
## Deliberately not built by this proof
|
|
- full treasury fee ledger
|
|
- generalized tax or accounting treatment
|
|
- multi-venue settlement harmonization
|
|
|
|
## Prevention requirements for this proof
|
|
- For every new `Completed`, `Not filled`, or realized attribution label:
|
|
- what exact durable record backs it
|
|
- what linkage key or heuristic connects it to the quote
|
|
- what uncertainty remains
|
|
- what regression test prevents future overclaiming
|